Modern Banking

[ 530 ]

M O D E R N B A N K I N G

Based on Italian data, Sapienza (1998) found loan rates declined when there was a substantial degree of market overlap and the market share of the acquired bank was small. However, if market share was large, loan rates increased.

Evenett (2003) takes a different approach. Using the database from the OECD (2001) study of consolidation of the financial sector in 13 industrialised countries, he tests for the effect of different types of M&As/strategic alliances on interest rate spreads. The spread is defined as the difference between the loan rate to prime rate customers less the deposit rate paid by commercial banks on demand, time or savings deposits.71 The OECD study distinguishes between domestic cross-border strategic alliances and M&As. In the estimating equation, the interest rate spread for each country is the dependent variable. Independent variables include the different types of strategic alliances and M&As, a fivefirm concentration ratio adjusted for the number of M&As in each country, the number of banks in the country, and control variables for regulatory changes and macroeconomic effects.72

Evenett finds a positive and significant coefficient on the concentration ratio: as concentration rises, so does the interest rate spread, which would suggest that any effects on efficiency brought about by the increase in consolidation have been more than offset by the effects of greater market power. The number of firms has a negative sign (consistent with the Cournot model), but is insignificant. Evenett also finds that domestic M&As raise spreads (coefficient significant and positive) but cross-border M&As do not (coefficient negative but insignificant). The coefficient on domestic alliances is also positive and significant, but for cross-border alliances, it is negatively signed and insignificant. Thus, increases in domestic consolidation and alliances raise spreads but their cross-border counterparts do not.

When the sample is split up and estimated by countries, the only notable change in results comes when Evenett divides the sample into the eight EU countries and the fivenation non-EU group. For the non-EU group, the coefficient on cross-border alliances is the only one that is significant – they reduce the interest rate spread. Cross-border M&As also have a negative coefficient, but it is insignificant. National alliances and M&As are positively signed but insignificant. By contrast, the EU group regression shows cross-border alliances raise spreads but cross-border M&As reduce them – the signs on the coefficients are respectively, positive and negative, and both are significant. Also, the EU regression has a significantly positive five-firm concentration ratio, and significantly negative coefficient on the number of banks.

Evenett’s results are interesting but should be treated with caution. Not only is the spread variable likely to be too aggregated and suffer from other problems (see footnote 69), but the controls for regulatory changes are somewhat arbitrary. The control variables do not include the ERM in the lead up to the adoption of the euro by most EU countries. Evenett’s results suggest that while efficiency gains appear to have outweighed market power

71Evenett (2003, p. 16) obtains these rates from the World Bank’s World Development Indicators. The definition alone is problematic. The ‘‘official’’ deposit and loan rates tend to be highly aggregated, e.g. an average across the main commercial banks. Furthermore, just one rate is provided for demand/time/savings deposits. Finally, the way rates are collected, ‘‘averaged’’, and reported varies among the national authorities.

72The macro control variables were real GDP and stock market capitalisation as a percentage of GDP in each country. Regulatory control variables were country-specific – see Appendix table 2 of Evenett (2003).

[ 532 ]

M O D E R N B A N K I N G

Ely and Song (2000) look at 60 US depository institutions with at least $5 billion in assets, operating from 1989 to 1995. These firms completed 449 acquisitions between 1990 and 1995. They test two hypotheses. According to the wealth maximisation hypothesis, in a period of regulatory reform, the most efficient banks will acquire the least efficient, which will create value and benefit the shareholders. By the incentive conflict hypothesis, managers acquire large numbers of banks to pursue their own interests.

A linear equation is estimated, where the dependent variable is the number of acquisitions (or the total assets gained through acquisitions) between 1990 and 1995. The independent variables are size (log of total assets), OWN (the percentage of shares held by insiders at year-end 1989) and the market to book value of equity (M/B), year-end 1989. If takeovers are shareholder wealth-maximising, the coefficient on M/B should be positive, if managers are pursuing their own interest, (M/B) will be negatively signed.

When the equation was estimated with acquired assets as the dependent variable, the wealth maximisation hypothesis is supported for the group of firms with or without an outside blockholder:74 acquisition activity is found to be positively (and significantly) related to the ratio of market to book value and to insider ownership. The positive coefficient on the insider ownership variable means the most active bidders are those which have minimised costs arising from agency conflicts. The group accounts for 67% of the acquired assets in the sample. However, if the other dependent variable (number of acquisitions) is used, the results are inconsistent with the wealth maximisation hypothesis and weakly support the incentive conflict hypothesis.75 Overall, shareholder value is more likely to be maximised if a few large acquisitions are made rather than a large number of small acquisitions. These findings are consistent with Bliss and Rosen (2001).

Hadlock et al. (1999) conduct an empirical test of the relationship between management incentives, performance and corporate governance in the target bank, and the likelihood of being acquired, using a sample of 84 acquired US commercial banks between 1982 and 1992, and another 84 that were not involved in a merger. They test a number of competing hypotheses.

žThe ‘‘irrelevance hypothesis’’: consolidation activities are not affected by management variables.

žThe ‘‘financial incentive hypothesis’’: managers with significant ownership of the bank make financial gains when their bank is acquired and therefore are more likely to be targets.

žThe ‘‘entrenchment hypothesis’’: managers who own a significant amount of the bank do not like to lose control and will reject attempts to be taken over, making continued independence likelier.

žThe ‘‘discipline hypothesis’’: poorly managed banks are more likely to be acquired. Better managed banks are those where managers have a significant ownership stake or there are

74A firm is deemed to have an outside blockholder if at least one outside investor owns at least 5% of the total shares.

75When the authors changed the dependent variables to year-end 1994, and the number of acquisitions is used as the dependent variable, the wealth maximisation hypothesis is supported, and accounts for about 70% of the 1995 acquisitions. No significant relationships are found if the dependent variable is acquired assets.

[ 533 ]

C O M P E T I T I V E I S S U E S I N B A N K I N G

a large number of outsiders on the board. Managers’ incentives are more in line with the shareholders’ incentives and/or the presence of outsiders in the board should ensure that actions are taken in the interest of shareholders. For these reasons, such banks are less likely to be targets.

Hadlock et al. find that banks with high levels of management ownership are less likely to be acquired. Furthermore, poor performance does not increase the likelihood of being acquired. They conclude the results are consistent with the ‘‘entrenchment hypothesis’’.

9.5.9. Consolidation and Systemic Risk

The subject of systemic risk was covered in Chapters 4 and 5. Mergers and acquisitions can affect systemic risk in a variety of ways. First, mergers and acquisitions generally result in a more diversified firm, which should reduce the risk. However, there is some research showing that the gains from diversification are offset by banks taking more risks, in search of greater expected returns. M&As also increase the size of banks and if these larger firms fail, the systemic consequences are normally greater. Regulators could find themselves with larger banks that come under the financial safety net of the central bank. If managers know a merged bank will be large enough to qualify for ‘‘too big to fail’’ status, they are more likely to take on greater risks, especially if the bank’s performance worsens – the ‘‘go for broke’’ syndrome. Thus, welfare costs could rise. The OECD (2001) report on consolidation among the G-10 nations found that the effects of M&As on banks’ risk are mixed, and that post-consolidation, banks may be inclined to take on more risks because they choose to or because monitoring of their risk exposure is less effective. The econometric research in this area has produced mixed results.

Berger et al. (1999), based on US evidence, found that BHCs move eligible activities from non-bank subsidiaries into bank subsidiaries. Kwast and Passmore (1998) provide the motivation for such action: BHCs are maximising shareholder value by ensuring that every activity they engage in can benefit from the implied subsidy of a financial safety net.

Benston et al. (1995) use US data to examine two competing hypotheses. The first is that banks merge to become ‘‘too big to fail’’. This behaviour would be in the interest of shareholders because it ensures all deposits are effectively protected by the financial safety net. Thus, they would expect the data to show that banks are willing to pay more for banks that will increase the post-merger bank risk.76 The second hypothesis posits that regulatory constraints mean increasing risk is not in the interest of shareholders. Instead, banks merge to diversify earnings and increase cash flow without increasing risk, so they are willing to pay higher prices for banks that will increase cash flow. The hypotheses are tested by looking at the relationship between the premium77 paid for the target bank and proxies for target’s and acquirer’s risk, risk diversification opportunities, the value of a deposit insurance put option, and the cash flow of the merged entity. Their findings are consistent with the second hypothesis, which also suggests that managers act in the interests of their shareholders.

76The preferred targets would be riskier banks whose returns are highly correlated with the acquirer bank.

77Measured as the price minus the preconsolidation market value of the target bank.

[ 534 ]

M O D E R N B A N K I N G

In a paper based on Italian bank mergers between 1997 and 2001, Chionsini et al. (2003) look at the effect of diversification on credit risk among merging banks. Instead of using the standard risk, return approach, these authors use data on the probabilities of default and loss given default for 180 000 non-financial Italian firms to assess the change in the credit risk profile of these banks before and after the mergers took place. They find that mergers significantly reduced credit risk because of the diversification of unsystematic risk. This result is likely due to the higher proportion of small and medium-sized enterprises (SMEs) in Italy. In their sample of 180 000 commercial and industrial firms, 80% were SMEs (sales of less than ¤5 million), accounting for 26% of loans.78 They argue SMEs are less likely to be affected by the economic cycle, and most of the default risk is firm or industry specific. The authors also studied the change in composition of the loan portfolios two years post-merger. They find no significant change in the banks’ portfolio risk, and loans to more creditworthy borrowers had increased. However, two years after the merger the findings indicated a larger diversification of systematic risk.

A paper by Acharya et al. (2002) also employed Italian data from 105 commercial banks in the period 1993 – 99. Though they do not look at the effects of M&As on diversification of risk, their findings are worth citing in light of the Chionsini et al. study. Using returns as the dependent variable, they test the effects of loan diversification on banks’ risk. For industrial and asset diversification,79 they find diversification simultaneously increases loan risk and reduces returns, implying an overall decline in bank performance. The opposite is the case if the diversification is geographic. Based on these results, Acharya et al. conclude that diversification can reduce returns in some cases, either because of poor monitoring and/or because of problems of adverse selection when banks attempt to move into new areas of activity.

9.5.10. Mergers and Acquisitions: A Case Study Approach

Davis (2000) takes a case by case approach on bank mergers and acquisitions, studying 33 mergers around the world that took place in the late 1990s. The discussion that follows embraces Davis’ key ideas but updates the information in the banks and provides some further analysis. The conclusions by Davis can be divided into several categories.

Strategic and financial targets

In this group of 33 banks, all had strategic targets, but just two-thirds of the sample gave explicit financial targets for the post-merger period. Many of these targets, such as reduced cost/income or increased ROE, may be the result of factors other than the merger itself, which is one of the problems with a case approach. Nonetheless, it is interesting to review Davis’ findings. Davis identified 15 that had been merged for long enough to comment on whether these targets have been met. Of the 15, six have achieved their financial targets.

78Firms with sales in excess of ¤50 million had 38% of total loans.

79Industrial diversification: expansion across different industrial sectors; asset diversification: expansion of loans across different asset classes.

[ 535 ]

C O M P E T I T I V E I S S U E S I N B A N K I N G

Banco Commercial Portuguesˆ (BCP) – Portugal: merged with Banco Portuguesˆ do Atlanticoˆ (BPA) in 1996.

Chase Manhattan: in 1991, Chemical merged with Manufacturers Hanover; in 1996, Chase merged with Chemical bank. Since then, Chase merged with JP Morgan, to form JP Morgan Chase in September 2000.

Den Danske Bank (DDB) – Denmark: formed in 1990, a three-way merger of Den Danske Bank, Copenhagen Handelsbank and Provinsbanken. It also became a regional Scandinavian Bank in the late 1990s, taking over Ostgoda Enskilda Bank (Sweden), Fokus Bank (Norway) and stockbrokers in Norway, Finland and Sweden.

Unibank – Denmark: after DDB, Denmark’s second largest bank. Again, it was the product of a three-way merger in 1990. In 1999, it acquired Tryg Baltica, the largest general insurance firm in Denmark.

Lloyds TSB: Now the fifth largest UK bank in terms of assets. Lloyds acquired the TSB group in December 1995, though full integration, in line with the regulatory authorities, did not take place until 1999. It also acquired Cheltenham and Gloucester (C&G), a former building society, but it has been kept as a separate subsidiary.

Svenska Handelsbanken: based in Sweden, this bank has retail operations in the other Nordic countries: Denmark, Finland and Norway. It took over a state owned mortgage bank, Stadshypotek, in February 1997.

The other nine banks failed to meet their financial targets, hence, the reasons for the outcomes are explored in more detail. The bank mergers include the following.

Bank Austria: took over Creditanstalt in 1997 by purchasing the government’s controlling share. Bank Austria had to agree not to change the structure of Creditanstalt for 5 years. As a result, there were two separate retail banking markets. The main reason for the agreement to keep Creditanstalt’s structure intact was concern that branches would be closed, causing reduced employment. This is reminiscent of the points made by Resti (1998) and Focarelli et al. (2000) – that inflexibility in some labour markets in European states can hinder gains from a merger. A single corporate structure BAC, it operates in global and investment banking. In 2004, it is still known as ‘‘BA-CA’’; its strategic goal is to become the leading bank in Central and Eastern Europe. Since 2000 it has been part of the German bank group HypoVereinsbank, which holds 22.5% of BA-CA’s shares.

HBV: Hypovereinsbank, Germany’s third largest bank, formed by the merger of two regional banks based in Munich: Bayerische Vereinsbank and Bayerische Hypotheken Bank, which was completed in October 1998. Dominant in mortgage markets in Europe and Germany, HBV was unable to meet its earnings targets because of the need to set aside DM 3.5 billion in provisions to cover losses from Hypo’s real estate lending in what was East Germany. As a result, all of the former Hypo executives resigned in October 1999. Hypovereinsbank is now the second largest private bank in Germany, with extensive operations in Austria (see above), Central Europe (with BA-CA) and Greece.

BBV: Banco Bilbao Vizcaya, Spain’s largest bank, formed in 1988 through the merger of two banks, roughly the same size. After the 1988 merger it invested in a new IT system which raised the cost to income ratio to between 60% and 70%, above the pre-merger level. It has since fallen to 56%. In 1999, it merged with the third largest bank, Argentaria, to become Banco Bilbao Vizcaya Argentaria (BBVA).

[ 537 ]

C O M P E T I T I V E I S S U E S I N B A N K I N G

banking system, and the cost income ratio target reduction has not been met because of an unexpected drop in revenues.

Falling stock prices

Most bank mergers have experienced a post-merger decline in the stock price, by more than 33% in some cases. Wells Fargo, First Union, Bank One, UBS and Swedbank have shown the most dramatic declines, for the reasons noted above. A weak share price can make a bank a target, as Wells Fargo found when it was taken over by Nor West. This is consistent with the findings of many academic studies, especially for US banks, that the stock price of the target bank falls. For bank mergers in their countries, see more recent data (e.g. Cybo-Ottone and Murgia, 2000; Beital et al., 2003; DeLong, 2003).

Unforeseen events

A common post-merger problem, unforeseen events suggest the need for careful due diligence before a takeover, to avoid finding post-merger skeletons in the cupboard. The problems encountered by those banks which failed to meet their financial targets provides a good example.

Politics

Politics can hamper mergers. Canada is the best example, where the government stopped the mergers of the Bank of Montreal with the Royal Bank and the Toronto Dominion Bank with Canadian Imperial Bank of Commerce in 1998. In Europe, unions, labour laws and general inflexibility can hamper the achievement of goals, especially when it comes to the closure of branches and staff reductions. In an econometric model using the time series data from 1976 to 1996, McIntosh (2002) shows the big five Canadian banks, with multiple branches, have technology characterised by increasing returns to scale, though the source of the scale efficiency is not determined. McIntosh (2002) speculates that it could arise because the cause of acquiring new technology is spread over a multibranch network. In simulations of what would have happened had the banks been allowed to merge, McKinnon shows that post-merger, ‘‘prices’’80 fall. He concludes that even though the mergers would have increased concentration, the scale efficiency offsets the effects of any reduced competition, and the post-merger price of banking services falls, and total banking services rise. McKinnon’s simulations show a social welfare gain, and the Canadian government should have allowed the mergers to take place. However, Baltazar and Santos (2003) use event study analysis, where the events are the announcement dates of the mergers and the data the government ruled the mergers would not be allowed. They infer81 that market power, rather than efficiencies from scale/scope economies, X-efficiencies or

80McIntosh uses a price index of banking services, but it is unclear where it comes from.

81Like many of these studies, there is no explicit test to confirm that investors reacted the way they did in anticipation of greater market power (and higher profits through higher prices).